We describe a method for using multiparameter flow cytometry to detect mitochondrial reactive oxygen species (ROS) in murine healthy hematopoietic stem and progenitor cells (HSPCs) and leukemia cells from a mouse model of acute myeloid leukemia (AML) driven by MLL-AF9.
We present a flow cytometric approach for analyzing mitochondrial ROS in various live bone marrow (BM)-derived stem and progenitor cell populations from healthy mice as well as mice with AML driven by MLL-AF9. Specifically, we describe a two-step cell staining process, whereby healthy or leukemia BM cells are first stained with a fluorogenic dye that detects mitochondrial superoxides, followed by staining with fluorochrome-linked monoclonal antibodies that are used to distinguish various healthy and malignant hematopoietic progenitor populations. We also provide a strategy for acquiring and analyzing the samples by flow cytometry. The entire protocol can be carried out in a timeframe as short as 3-4 h. We also highlight the key variables to consider as well as the advantages and limitations of monitoring ROS production in the mitochondrial compartment of live hematopoietic and leukemia stem and progenitor subpopulations using fluorogenic dyes by flow cytometry. Furthermore, we present data that mitochondrial ROS abundance varies among distinct healthy HSPC sub-populations and leukemia progenitors and discuss the possible applications of this technique in hematologic research.
Reactive Oxygen Species (ROS) are highly reactive molecules derived from molecular oxygen. The most well-defined cellular location of ROS production is the mitochondria, where electrons that pass through the electron transport chain (ETC) during oxidative phosphorylation (OXPHOS) are absorbed by molecular oxygen leading to the formation of a specific type of ROS called superoxides1. Through the actions of a series of enzymes, called superoxide dismutases or SODs, superoxides are converted into hydrogen peroxides, which are subsequently neutralized into water by enzymes such as catalase or glutathione peroxidases (GPX). Perturbations in ROS-regulatory mechanisms can lead to the excess production of ROS, often referred to as oxidative stress, which have harmful and potentially lethal cellular consequences such as macromolecule damage (i.e., DNA, protein, lipids). Moreover, oxidative stress is related to several pathologies, such as diabetes, inflammatory diseases, aging and tumors2,3,4. To maintain redox homeostasis and prevent oxidative stress, cells possess a variety of ROS-regulating mechanisms5.
Physiological levels of certain ROS are necessary for proper embryonic and adult hematopoiesis6. However, excess ROS is associated with DNA damage, cellular differentiation and exhaustion of the hematopoietic stem and progenitor pool. There is also evidence that alterations in redox biology may differ between leukemia and healthy cells. For example, ROS levels tend to be higher in acute myeloid leukemia (AML) cells relative to their healthy counterparts and other studies have suggested that leukemia stem cells maintain a low steady-state level of ROS for survival7,8. Importantly, strategies for therapeutically capitalizing on these redox differences have shown promise in several human cancer settings9,10. Therefore, assays that allow for the assessment of ROS levels in mouse models may improve our understanding of how these species contribute to cellular physiology and disease pathogenesis as well as potentially provide a platform for assessing the effectiveness of novel redox-targeting anti-cancer therapies.
All of the animal procedures described in this protocol have been approved by the Institutional Animal Care and Use committee (IACUC) at Fox Chase Cancer Center.
NOTE: The protocol workflow is divided into 4 parts as presented in Figure 1 and the required reagents are listed in the Table of Materials.
1. Bone Marrow (BM) Isolation
NOTE: MLL-AF9 leukemia mice were generated as described previously11.
2. Mitochondrial ROS Fluorogenic Dye Staining
3. Lineage Antibody Staining
4. Flow Cytometry Acquisition and Analysis
NOTE: Several hematopoietic stem and progenitor subsets are rare, such as long-term hematopoietic stem cells. Thus, ideally 3-5 million events should be collected for each experimental tube during flow cytometry acquisition for sufficient analysis of mitochondrial ROS in the various HSPC subsets.
Presented is a method for analyzing ROS in the mitochondria of multiple healthy and MLL-AF9-expressing leukemia progenitor populations. Figure 1 displays a schematic view of the protocol workflow, which consists of 4 major steps: 1) BM isolation from mice; 2) Staining BM cells with a fluorogenic dye that recognizes mitochondrial ROS, particularly superoxides; 3) Surface marker antibody staining to delineate various healthy and leukemia hematopoietic populations; and 4) Flow cytometry acquisition and analysis.
Figure 2A,B depicts a representative gating strategy for analyzing various hematopoietic stem and progenitor populations in healthy and leukemia BM. An FSC/SSC plot is applied to eliminate debris and an FSC-area (FSC-A) versus FSC-height (FSC-H) plot is used to exclude doublets and aggregates. A panel of lineage surface markers (Table 1 and step 3.1) are combined to exclude a variety of mature hematopoietic populations such as lymphocytes, erythrocytes, granulocytes and monocytes/macrophages (i.e., Lineage low or Linlow). The lineage cocktail also includes a CD48 antibody and therefore all subsets of lineage low cells are also CD48–. Sca-1 and c-Kit cell surface expression is used to distinguish a heterogeneous mixture of HSPCs called CD48– LSKs (Linlow, Sca-1+, c-Kit+, CD48–) from myeloid progenitors (Linlow, Sca-1–, c-Kit+, CD48–). To further distinguish various HSPC subsets, the SLAM marker, CD150 as well as CD34 are also added17,18,19,20,21. Figure 2B also depicts a representative gating strategy for cKit high-expressing leukemia progenitors (Linlow, c-Kithigh; Sca-1–), which are enriched for leukemia initiating cells (LICs)11 as well as leukemia cells expressing intermediate-low expression of cKit (cKitInt-low).
A comparison of mitochondrial ROS staining between healthy CD48– LSK and myeloid progenitors shows that myeloid progenitors display significantly higher levels of mitochondrial ROS staining (Figure 3A). Moreover, cKithigh leukemia progenitors display significantly higher levels of mitochondrial ROS staining compared to CD48– LSK, myeloid progenitors or cKitint-low leukemia cells (Figure 3A). cKitInt-low leukemia cells also displayed significantly higher mitochondrial ROS staining compared to CD48– LSK cells but not to myeloid progenitors (Figure 3A). LSKs further sub-divided by CD150 expression showed that mitochondrial ROS staining did not significantly vary in CD48– LSK cells sub-divided by CD150 high (CD150High), intermediate (CD150Int) or no (CD150Neg) expression (Figure 3B). However, steady-state mitochondrial ROS staining of cKithigh or cKitInt-low leukemia cells was found to be significantly higher than CD48- LSKs-CD150High, -CD150Int or -CD150Neg cells (Figure 3B). LSK cells expressing low to no levels of CD34 are enriched for long-term reconstituting hematopoietic stem cells, particularly CD48- LSK cells that are CD34– and CD150high 20,21. However, subdividing CD48– LSK cells by CD150 and CD34 expression did not reveal any significant differences in mitochondrial ROS staining amongst these six HSPC subsets (Figure 3C).
Figure 1: Schematic representation of the protocol work-flow. Step 1) BM isolation from healthy and MLL-AF9 leukemic mice; Step 2) cellular staining using a mitochondrial ROS dye (mROS); Step 3) cellular staining with fluorochrome-linked monoclonal antibodies to discriminate hematopoietic stem and progenitor cells (HSPCs) population in healthy and leukemic mice; Step 4) Flow Cytometry acquisition and analysis of mitochondrial ROS in several HSPC populations. Please click here to view a larger version of this figure.
Figure 2: Flow cytometry gating strategies for healthy and MLL-AF9-expressing bone marrow cells. (A) BM cells isolated from healthy mice were stained with a live/dead dye (QDot), mitochondrial ROS dye (TRPE). BM from healthy mice was subsequently stained with antibodies recognizing lineage markers plus CD48 (Cy5-PE), c-Kit (Cy7-APC), Sca1 (PacBlue), CD34 (FITC), CD150 (APC). (B) In addition to live/dead cell and mitochondrial ROS stains, BM from leukemia mice were also stained with antibodies recognizing lineage markers plus CD48 (Cy5-PE), c-Kit (Cy7-APC), Sca1 (PacBlue) and CD45.2 (APC), which is applied to discriminate between MLL-AF9 leukemia cells from healthy recipient BM cells (CD45.1). Please click here to view a larger version of this figure.
Figure 3: Mitochondrial ROS levels in healthy and MLL-AF9 Bone Marrow. Left panels are representative histograms of mitochondrial ROS levels of the indicated populations and the respective right panels represent bar graph analyses of the MFI of mitochondrial ROS for the indicated populations (n = 4). (A) Comparison of mitochondrial ROS levels in healthy CD48– LSK, myeloid progenitors as well as MLL-AF9 Linlow c-KitHigh and MLL-AF9 Linlow c-KitInt-Low cells. (B) Comparison of mitochondrial ROS levels in healthy CD48– LSK cells based on their CD150 expression versus MLL-AF9 Linlow c-KitHigh and MLL-AF9 Linlow c-KitInt-Low cells. (C) Comparison of mitochondrial ROS levels in healthy CD48– LSK cells based on their CD150 and CD34 expression. (* p≤0.05; ** p < 0.01*** p < 0.001; **** p < 0.0001). Please click here to view a larger version of this figure.
Figure 4: Changes in mitochondrial ROS levels using pro- and anti-oxidant compounds. Histograms of mitochondrial ROS levels in healthy and MLL-AF9-expressing BM cells treated with 20 µM of menadione sodium bisulfite (MSB) for 1 h at 37 ˚C in a 5% CO2 incubator (positive Control) or with 20 µM of MSB in combination with 100 µM N-acetyl-L-cysteine (NAC) for 1 h at 37 ˚C in a 5% CO2 incubator. Please click here to view a larger version of this figure.
Figure 5: Optimization of the mitochondrial ROS and lineage-recognizing antibody stains. (A) Comparison of mitochondrial ROS (mROS) levels in MLL-AF9 expressing leukemia cells using 1 or 5 µM for either 10 or 30 min. (Red = vehicle; Blue = MSB 20 µM; Green = NAC 100 µM; Orange = Msb 20 µM + NAC 100 µM). (B) Comparison of the MFI of the CD34 channel (FITC) in healthy LSKs stained for 20, 60 or 90 min with the antibody cocktail #1 (n = 4, * p ≤ 0.05; ** p < 0.01). (C) Comparison of antibodies staining before or after incubation for 30 min at 37 ˚C in a 5% CO2 incubator. Please click here to view a larger version of this figure.
Figure 6: Order of mitochondrial ROS and lineage marker antibody staining. (A) Dot plots of the indicated HSPCs populations obtained by using different orders of staining. (B) Mitochondrial ROS levels evaluated in the LSK and Myeloid Progenitors compartments obtained used different order of staining (Red = antibody staining for 1 h at 4 °C followed by mitochondrial ROS staining for 10 min at 37 °C; Blue = mitochondrial ROS staining for 10 min at 37 °C followed by antibody staining 1h at 4 °C). (C) Quantification of the MFI of mitochondrial ROS (mROS) staining using different orders of staining in the indicated populations (n = 4, * p ≤ 0.05). Please click here to view a larger version of this figure.
Figure 7: Impact of verapamil treatment on mitochondrial ROS dye staining in healthy and MLL-AF9-expressing BM cells. (A) Dot plots of the indicated HSPC populations in samples treated with or without 50 µM Verapamil for 10 min at 37 °C in a 5% CO2 incubator. (B) Histograms of mitochondrial ROS and mitoMASS Green dye staining in the indicated HSPC populations in the presence (blue) or absence (red) of 50 µM Verapamil. (C-E) Quantification of mitochondrial ROS levels in the indicated healthy (C & D) and leukemia (E) cell populations in BM samples treated with or without 50 µM Verapamil during mitochondrial ROS staining (n = 4, * p ≤ 0.05). Please click here to view a larger version of this figure.
Table 1: Antibody cocktails. List of antibody cocktails prepared in Step 3.1. to identify various hematopoietic sub-populations within healthy and leukemia bone marrow.
Fluorogenic dyes that have been developed for the detection of ROS are frequently evaluated in fixed cells by microscopy or in live cells by flow cytometry22. Flow cytometric evaluation of mitochondrial ROS in BM cells using mitochondrial ROS fluorogenic dyes has two major advantages: 1) It is a fast and simple technique that is suitable for live cell analysis and 2) it allows for distinguishing and analyzing rare populations at the single-cell level in the BM using surface marker staining. The step-by-step protocol presented here has been developed to study the ex vivo redox status of hematopoietic stem and progenitor populations from both healthy mice as well as a mouse model of AML driven by MLL-AF9 using flow cytometry. There are several key technical variables that need to be considered during the execution of this protocol.
First, the use of pro- and anti-oxidant controls allows the user to establish a baseline for increases in mitochondrial ROS staining as well as the specificity of the stain. For the protocol presented here, the pro-oxidant MSB was utilized as a positive control for a detectable induction of mitochondrial ROS staining in both healthy and leukemia BM cells (Figure 4). The anti-oxidant NAC can be used as an additional control, as it largely reverses the mitochondrial ROS staining induced by MSB (Figure 4).
Second, the mitochondrial ROS fluorogenic dye employed in this study is recommended to be used at a concentration of 5 µM for 10 min. However, the manufacturer also suggests that the concentration and time of staining may vary between cell types. In this study, mitochondrial ROS staining was compared at both 1 µM and 5 µM for either 10 or 30 min. The analysis revealed that a concentration of 5 µM for either 10 to 30 min is sufficient to detect MSB-mediated changes in mitochondrial ROS as well as those reversed by NAC treatment (Figure 5A). Since there was no quantitative difference between 10 and 30 min, a staining time of 10 min was selected for this study to minimize the length of the assay.
Third, CD34 antibody incubation times vary within the literature from 20 to 90 min23,24. To optimize the protocol presented here, murine bone marrow cells were incubated with the antibody cocktail #1 (Step 3.1) for 20, 60 or 90 min. As shown in Figure 5B, significantly stronger CD34 staining was observed on cells stained for 60 or 90 min compared with the 20 min stain. However, a significant difference in CD34 staining was not observed between antibody incubation times of 60 and 90 min (Figure 5B) and thus a 60 min antibody stain is recommended for the presented protocol.
Fourth, in the presented protocol, both healthy and leukemia cells are first stained with the mitochondrial ROS fluorogenic dye, washed and then stained with fluorochrome-linked lineage antibodies. Although a direct assessment of combining the mitochondrial ROS stain with lineage antibodies was not conducted in this study, an evaluation of fluorochrome-linked lineage antibody staining was assayed under mitochondrial ROS staining conditions for 10, 20 and 30 min at 37 °C. This analysis revealed that 30 min of incubation at 37 °C substantially alters lineage surface marker staining (Figure 5C). Additionally, mitochondrial ROS staining was evaluated by first, staining healthy BM cells with fluorochrome-linked lineage antibodies followed by staining with the mitochondrial ROS fluorogenic dye as well as vice versa order of operations. Staining cells first with lineage antibodies followed by mitochondrial ROS staining resulted in lower mitochondrial ROS signals compared to the vice versa staining protocol (Figure 6A,B) - including significant differences for CD48– LSK CD150high, CD48– LSK CD150Int CD34 and myeloid progenitor populations (Figure 6C).
Fifth, recent studies show that different murine healthy HSPC populations possess distinct abilities to efflux several mitochondrial staining fluorogenic probes such as those used to assess mitochondrial mass (hereafter referred to as mitoMASS green) or potential25,26. Therefore, the impact of the efflux pump inhibitor verapamil on the staining of healthy and leukemia progenitors with the mitochondrial ROS fluorogenic dye was also assessed. Simultaneous staining of HSPCs with mitochondrial ROS did not alter lineage marker antibody staining (Figure 7A), however, verapamil did significantly improve mitochondrial ROS staining signals in a variety of healthy and leukemia progenitor populations (Figure 7B,C). Notably, the magnitude of improved mitochondrial ROS staining by verapamil was not as great as that seen with the mitoMASS green fluorogenic dye (Figure 7B).
Sixth, in the protocol presented here, BM cells were recovered by removing the bone tips (epiphysis) followed by flushing of the bones. However, the epiphysis contains hematopoietic cells that could potentially be lost by bone flushing. As an alternative, BM cells can be extracted by mashing bones with a mortar and pestle as previously described12,13.
The protocol presented here provides a foundation for the use of fluorogenic dyes to measure intracellular redox biology in live cells extracted from living organisms. However, it is important to note that no single fluorogenic dye can be assumed to be definitively specific and therefore additional studies with alternative methods should be conducted to verify any findings. Furthermore, the results of this study suggest that leukemia cell populations enriched for LICs (i.e., Linlow cKithigh) display higher levels of mitochondrial ROS than healthy myeloid progenitors or other HSPC populations. However, the Linlow cKithigh leukemia cell population evaluated here has been shown to be heterogeneous and can be further sub-divided by other lineage markers11,27. Furthermore, recent studies show that LICs possess a distinct metabolic phenotype28. Therefore, future studies assessing mitochondrial ROS in parallel with metabolic assays or probes as well as additional lineage marker antibodies will be informative.
This straightforward protocol allows for the measurement of mitochondrial ROS levels in living hematopoietic cells and may provide a basis for studying the redox biology of healthy and diseased stem and progenitor cells as well as for assessing the effectiveness of redox-targeting therapies.
The authors have nothing to disclose.
This work was supported by The Fox Chase Cancer Center Board of Directors (DDM), the American Society of Hematology Scholar Award (SMS), American Cancer Society RSG (SMS) and the Department of Defense (Award#: W81XWH-18-1-0472).
Heat inactivated FBS | VWR Seradigm LIFE SCIENCE | 97068-085 | Media |
Penicillin Streptomycin | Corning | 30-002-CI | Media |
PBS | Fisher Scientific | BP399-20 | Buffer |
15 mL conical tube | BD falcon | 352096 | Tissue Culture Supplies |
50 mL conical tube | BD falcon | 352098 | Tissue Culture Supplies |
40 μm cell strainers | Fisher Scientific | 22-363-547 | Tissue Culture Supplies |
RBC Lysis Buffer | Fisher Scientific | 50-112-9751 | Tissue Culture Supplies |
Menadione sodium Bisulfite | Sigma aldrich | M5750 | Pro-oxidant |
NAC | Sigma aldrich | A7250 | Anti-oxidant |
CD3 PE-Cy5 clone 145-2c11 | Biolegend | 100310 | Antibody |
CD4 PE-Cy5 clone RM4-5 | eBioscience | 15-0041-81 | Antibody |
CD8 PE-Cy5 clone 53-6.7 | eBioscience | 15-0081-81 | Antibody |
CD19 PE-Cy5 clone 6D5 | Biolegend | 115510 | Antibody |
B220 PE-Cy5 clone RA3-6B2 | Biolegend | 103210 | Antibody |
Gr1 PE-Cy5 clone RB6-8C5 | Biolegend | 108410 | Antibody |
Ter119 PE-Cy5 clone Ter-119 | Biolegend | 116210 | Antibody |
CD48 PE-Cy5 clone HM48-1 | Biolegend | 103420 | Antibody |
CD117 APC-Cy7 clone 2B8 | Biolegend | 105825 | Antibody |
Sca1 peacific Blue clone D7 | Biolegend | 108120 | Antibody |
CD150 APC clone TC15-12F12.2 | Biolegend | 115909 | Antibody |
CD34 FITC clone RAM34 | BD Bioscience | 553733 | Antibody |
CD45.2 APC clone 104 | Biolegend | 1098313 | Antibody |
MitoSOX Red | ThermoFisher Scientific | M36008 | Dye |
Mitotracker Green | ThermoFisher Scientific | M7514 | Dye |
Live/dead Yellow Dye | ThermoFisher Scientific | L34967 | Dye |